Emulsion polymerization of an epoxy-acrylate emulsion stabilized with polyacrylate. i. influence of...

Emulsion Polymerization of an Epoxy-Acrylate EmulsionStabilized with Polyacrylate. I. Influence of Salt, Initiator,Neutralizing Amine, and Stirring Speed

D. DE WET-ROOS,1 J. H. KNOETZE,2 B. COORAY,3 R. D. SANDERSON1

1 Institute of Polymer Science, University of Stellenbosch, Private Bag X1, Matieland, Stellenbosch, 7602,Republic of South Africa

2 Department of Chemical Engineering, University of Stellenbosch, Private Bag X1, Matieland, Stellenbosch, 7602,Republic of South Africa

3 Plascon Pty. Ltd., Group Head Office, P.O. Box 1227, Johannesburg, 2000, Republic of South Africa

Received 20 October 1997; accepted 17 August 1998

ABSTRACT: The past decade has seen the development of high-performance epoxy-acrylate coatings. Some of these coatings are used exclusively as can coatings. Toimprove chemical resistance, emulsion polymerization in the presence of the dispersedepoxy resin can be conducted. Replicated factorial designs were designed to investigatesome of the factors that contribute to the performance and stability of the epoxy-acrylate coating. The factors investigated in this study were: the type and amount ofneutralizing amine, the type of initiator, the presence or absence of salt, and thestirring speed of the impeller. The measured responses for the experimental designwere particle size, particle size distribution, and conversion. It has been shown in thisstudy that the type of amine used to neutralize the polyacrylic stabilizer significantlyinfluenced the observed responses probably by acting as chain transfer agent duringpolymerization. The addition of salt significantly affected the observed responses byshielding the surface charges on the particle surface that impart stability to the latexthrough electrostatic repulsion. Stirring speed of the impeller and the type of initiatoras factors in the range investigated do not influence the kinetics or the particle size andpolydispersity to any great extent either on their own or in combination with otherfactors. © 1999 John Wiley & Sons, Inc. J Appl Polym Sci 71: 1347–1360, 1999

Key words: epoxy-acrylate; emulsion polymerization; can coating; statistically de-signed experiments; conversion; particle size; polydispersity

INTRODUCTION

During the past 10 years, legislation in theUnited States and Europe on the reduction of thevolatile organic content of paints and coatings hasresulted in the development of many new water-

reducible coatings. Among these is the water-re-ducible epoxy coatings.1–11

Research into these water-based epoxy coat-ings has resulted in the development of so-calledtransitional products, so named because they sig-nified the transition of the traditional solvent-based coatings to coatings dispersed in water. Inthese transitional water-based coating systems,low molecular weight diglycidyl ethers of bisphe-nol A were emulsified with combinations of an-

Correspondence to: D. de Wet-Roos.Journal of Applied Polymer Science, Vol. 71, 1347–1360 (1999)© 1999 John Wiley & Sons, Inc. CCC 0021-8995/99/081347-14

1347

ionic and nonionic surfactants.11 The develop-ment of products with even lower volatile organiccontent contents than the previous transitionalproducts and the use of acid-salted amine adductsto impart dispersion of the epoxy in water fol-lowed. These water-based epoxy coatings weretermed “first-generation coating systems.”

In these first-generation coating systems, hy-drophilic molecules were incorporated into thechemical structure of the epoxy to make it water-dispersible.1–6,10 Grafting was extensively used toproduce water-reducible epoxy graft copoly-mers.4,12

In an example of a grafting reaction, additionpolymerization of combinations of hydrophilicand hydrophobic monomers initiated by a stan-dard initiator, such as benzoyl peroxide, was con-ducted in the presence of an epoxy resin.4 Afterpolymerization, the grafted epoxy resin, alongwith a significant amount of ungrafted copolymer,was dispersed as micelles in the water phase byneutralizing the carboxylic acid groups with asuitable amine. The main problem associatedwith most of these first-generation products wasthe unsatisfactory chemical and water resistance,combined with their somewhat inferior film-form-ing ability and the inability to be applied to me-tallic substrates.

In the late 1980’s and early 1990’s, the follow-ing observations were made regarding high-per-formance coatings based on acrylic dispersants orpolyelectrolytes. Much higher molecular mass ep-oxies, compared with previous water-based sys-tems with a resultant increase in coating quality,could now be formulated. The resultant epoxy-acrylate emulsion was also able to undergo self-curing after application as a coating on a sub-strate.12–14 Many of these latest epoxy-acrylatewater-based emulsions are used commercially asmetal coatings with an improved film-formingability and for high-gloss applications. There isalso some epoxy-acrylate emulsions that are de-signed specifically to be applied as can coatings.14

Can coatings for alcoholic beverages are re-quired to have superior film-formation ability, bedurable, and exhibit high chemical and water re-sistance, as well as possess specific rheologicalproperties.

In our laboratories, we are studying ways toincrease the chemical and water-resistance of cancoatings. This is achieved by adding combinationsof suitable monomers, such as butyl acrylate andstyrene to the predispersed epoxy-acrylate disper-

sion and initiating the polymerization with a suit-able initiator, such as ammonium persulfate.

The aim of this article is to highlight the effectsthat various factors that contribute to a stable,high-performance epoxy-acrylate emulsion haveon the overall conversion of monomers (such asbutyl acrylate and styrene to the resultant copol-ymer, particle size, and polydispersity of theemulsion). Examples of these factors include stir-ring speed of the agitator, the presence of dis-solved salt in the water, the type of initiator usedfor addition polymerization, and the amount andtype of amine used to neutralize the polyacrylate.

BACKGROUND

The epoxy-acrylate dispersion (dispersed epoxyphase) and later the epoxy-acrylate emulsion(with added monomer) are stabilized mainly bythe contribution of the neutralized carboxylic acidgroups incorporated during the synthesis of thepolyacrylate. The number of ionic groups on thesurface of the particle influences the thickness ofthe diffuse electric layer around the particle. Thethickness and therefore the stability of the dis-persed particle are influenced by the thickness ofthis layer. The thickness is given in terms of theDebye length 1/k. k is given by the following equa-tion:

k 5 S8pnZ2e2

«kT D (1)

In this equation, the symbol n denotes the num-ber of ionized groups, Z is the valence of theionized groups, e is the elementary charge, « isthe dielectric constant, k is the Boltzmann con-stant, and T is the temperature in Kelvin.

An initiator, such as ammonium persulfate,can contribute to the latex stabilization throughthe ionic end-groups of the polymer chains. Apartfrom electrostatic stability, a degree of steric sta-bilization may also contribute to the overall sta-bility of the latex. This may be possible throughthe extension of parts of the polyacrylate contain-ing enough ionic carboxylic acid groups extendinginto the medium as long chains or loops. However,because these chains also contain ionized groups,they will be susceptible to the same factors thatdestabilize the charges on the surface of the par-ticle.

1348 DE WET-ROOS ET AL.

The difference between coagulation and floccu-lation needs to be qualified because these twoterms are sometimes used synonymously. Coagu-lation occurs when the forces holding particles insuspension are overcome or neutralized. Floccu-lation on the other hand occurs when groups ofparticles aggregate into larger groups of particlesor flocs.

Coagulation may be described as the result oftwo or more particles colliding with enough forceto break through the electrostatic barrier andsteric hindrance surrounding the latex particle toform a new larger particle: (1) by reducing thepotential energy barrier due to the electrostaticand steric stabilization can accelerate coagulationand (2) by increasing the probability of a collisionencounter between particles.

Simple electrolytes, such as NaCl in water, actby shielding the charges on the surface of thelatex particle and compressing the electric doublelayer surrounding the particle.15–17 In this way,the thickness of the diffuse electric layer is re-duced and, in so doing, the potential energy bar-rier to coagulation is lowered. Reduction of repul-sive forces allows the attractive forces betweenparticles to become dominant.

The kinetics of coagulation assume that theprocess is diffusion-driven.17 This means that astable colloidal system of uniform primary parti-cle size that is suddenly subjected to conditionsthat will enhance flocculation and coagulationwill experience diffusion of particles toward a pri-mary particle that will act as a center for furtheragglomeration. This theory also assumes that allparticle collisions are effective in causing floccu-lation. However, the presence of a potential bar-rier surrounding the particles will have the effectthat only a certain percentage of particle colli-sions will result in coagulation.

Agitation will promote the probability of colli-sions and therefore coagulation. This probabilitytakes into account the velocity gradient that theparticle is subjected to according to the followingequation:

J 5 ~43 N~Rij!

3~du/dz! (2)

The probability of collision of particle i with par-ticle j is given by J. Rij is the collision radius, anddu/dz is the velocity gradient. The combination ofan electrolyte and increased rate of agitation willenhance the probability of collision and eventualcoagulation.

If the number of loci or particles change signif-icantly with time during an emulsion polymeriza-tion, this will also affect the rate of polymeriza-tion and the extent of conversion of monomer topolymer. There is a direct dependence of the rateof emulsion polymerization on the number of locifor polymerization (Rp } Nc). This becomes clearif one examines the rate equation for emulsionpolymerization.18

Rp 5dxdt 5

Cpkpn# Nc

nm0 NA

(3)

In eq. (3), Cp is the monomer concentration insidethe particle, kp is the propagation rate constant, nis then the average number of radicals per parti-cle, Nc is the number of latex particles per dm3,nm

0 is the starting monomer concentration perdm3 latex, and NA is Avogadro’s number.

It is also well known that ionic strength of themedium affects the mass transfer of monomerand radicals in conventional emulsion polymer-ization. It is therefore important to understandthe effects that the various factors discussed inthis study will have on the stability and mecha-nism of polymerization.

EXPERIMENTAL

Synthesis of Polyelectrolyte

The polyfunctional acrylic polymer (polyacrylate)was synthesized by charging 1,755 g of butyl cel-losolve to a 10 L stainless steel Buchi batch reac-tor. The reactor was heated to a temperature of130°C and then 545 g of methacrylic acid, 563.8 gof 2-hydroxyethyl acrylate, and 2,112 g of butylacrylate were fed to the reactor, according to afixed monomer feed regime. At the same time,benzoyl peroxide, dissolved in 550 mL butyl cel-losolve, was added to the reactor at a prescribedrate. Changing the initiator concentration, aswell as the addition rates of monomer and initia-tor, could vary the molecular mass of the result-ant acrylic polymer. The polyacrylate was used,as it was prepared without further purification.The solids content of the solution was determinedby drying a weighed sample for 2 h at 150°C in anoven equipped with a fan.

Experimental Design

Two types of initiator were investigated for theemulsion copolymerization of styrene and butyl ac-

EMULSION POLYMERIZATION OF AN EPOXY-ACRYLATE EMULSION 1349

rylate, namely 2,29-azo-isobis-butyronitrile (AIBN),recrystallized from methanol, and ammonium per-sulfate purchased from Fluka (Ronkonkoma, NY)and used as received.

The emulsion polymerization experiments forthe determination of the effects of agitator stir-ring speed, initiator type, neutralizing aminetype, and concentration, as well as the effect ofdissolved salt on fractional conversion, particlesize, and polydispersity of the emulsions, weredesigned as center-point replicated factorial de-signs, with the factors and levels investigated foreach initiator type given in Table I. Polydispersityin the context of this study is an estimate of thevariance of distribution. This is a dimensionlessindex of the broadness of the particle size distri-bution.

The responses measured were Z-average par-ticle size and polydispersity at the end of emul-sion polymerization and fractional conversion af-ter 6,000 s of the total monomer concentration.The standard recipe for emulsion polymerizationis summarized in Table II.

In the future, references to low- and high-factorlevels will refer to the amounts given in Table I.The results of the factorial designed experimentswill be illustrated in terms of response surfacegraphs. These graphs are the product of multipleregression models of the main factors and theirinteractions.19,20 The multiple linear regressionmodels have the following general formula for adesign in b replicated blocks:

Yu 5 b0 1 Oi 5 1

k

bixui 1 Oi 5 1

k

biixi2

1 O Oi , j

k

bijxuixuj 1 Ol 5 1

b

dlzul 1 « (4)

In eq. (4), Yu denotes the observed response valueof the uth experimental run, xui is the corre-sponding ith input variable, dl denotes the effect

of the lth block, zul is a dummy variable havingthe value of 1 if the uth trial is conducted in thelth block. If this is not the case, zul is 0. In eq. (1),b0, b1, . . . , bk are the unknown parameters forwhich estimates are sought to fit a model to thephysical data.

The molecular mass of the electrosteric polyac-rylate used to disperse the epoxy in the waterphase was kept constant throughout. The polyac-rylate solution had a solids content of 58%. Theepoxy used was SHELL’s EPIKOTE 3001. Thesolid epoxy was dissolved in butoxy ethanol toyield a solution with a total solids content of 73%.Monomers were washed with 0.3M NaOH to re-move as much inhibitor as possible and then dis-tilled under reduced pressure. In all instances,the middle fraction of the distillate was kept.Monomers were stored at 210°C and were usedwithin 1 week after distillation.

Emulsion Polymerization

In a typical emulsion polymerization experiment,the correct amounts of epoxy resin solution andpolyacrylate solution were weighed off and mixedtogether. Dimethyl aminoethanol (DMEA) and asmall amount of water were then added to theepoxy/polyacrylate mixture to yield a water-in-oilemulsion. This water-in-oil emulsion was in-verted to give an oil-in-water emulsion by theaddition of the remainder of the water as stated inTable II. Depending on the requirements of theexperimental design, salt was added (see Table I).The pH was measured with a pH meter bothbefore and after polymerization.

The emulsion was degassed for 1 h beforepolymerization and also during polymerization,

Table II Standard Recipe Used for Epoxy-Acrylate Emulsion Polymerization

Ingredients Quantities

Surfactant batch no. Apt4r1/Apr1Surfactant amount 70 gStyrene amount 20 gButyl acrylate amount 20 gEpoxy amount 20 gInitiator amount 4.3821 E-03 molDMEA amount VariableSodium chloride amount VariableDistilled deionized water 350 gTemperature 70°CStir speed Variable

Table I Upper and Lower Levels of Factors forFactorial Designed Experiments

Factor Lower Level Upper Level

DMEA level 10 g 40 gStirspeed 100 rpm 600 rpmSodium chloride level 0 g 5 g


by bubbling nitrogen gas through the emulsion.Before it was decided to bubble nitrogen throughthe emulsion, experiments were done to ensurethat this did not cause surface coagulation. Afterthe initial degassing period, monomer as well asinitiator were added to the reactor. The timewhen the monomer was added to the emulsionwas taken as the beginning of the experiment. Allemulsion copolymerizations were conducted overa period of 5 h (18,000 s) in a water bath regulatedto a temperature of 70°C.

Samples of the emulsion were taken at regularintervals and analyzed by means of gas chroma-tography (GC) for monomer conversion. Particlesizes (Z-average) and polydispersity measure-ments were obtained with a Malvern Zetasizer4S. The GC used was a Perkin-Elmer AutosystemGC (Perkin-Elmer, Norwalk, CT). The GC wasequipped with a capillary column of 45 m inlength and 0.32 mm in diameter. The stationaryphase was 2 mm thick and was based on poly(m-ethyl silicone). The detector was a FID detector.

Chromatograms were obtained and integrated bymeans of EZCHROM software on a personal com-puter linked to the GC.

Sample Handling

Samples for GC analysis were withdrawn fromthe reactor during runs with a 1-mL syringe atregular intervals. All samples were quickly addedto a vile containing a solution of hydroquinone,acetic acid, and butoxy ethanol. The reason forthe acetic acid was to adjust the pH for the hy-droquinone to be effective as inhibitor. Directlyafter addition of the withdrawn sample to thequenching solution, the vial was cooled down inice. Dimethylformamide was added as internalstandard to every sample taken.

Samples for analysis by the Malvern particlesize analyzer were withdrawn at regular inter-vals during emulsion polymerization. Samplingconsisted of adding 2–5 drops of the emulsion to avial containing a solution of deionized water con-

Table III Estimates of Coefficients for Multiple Regression Model

Input Variables

Observed Response

Particle Size Polydispersity Conversion

A. Stirring Speed 0.104 23.62E-04 1.04E-4B. DMEA 5.28 20.016 20.016C. Salt 100.38 20.168 20.168AB 20.0037 1.8E-7 1.8E-7AC 20.00431 2.7E-5 2.7E-5BC 22.339 3.2E-3 3.2E-3Block 37.11 20.035 20.0093Constant 72.61 0.40 0.887

Figure 1 Response surface graphs showing the effects of stirring speed, DMEAconcentration, and the addition of salt on particle size of the emulsion after 5 h ofemulsion polymerization.


taining 1 mM NaCl. The exact amount of latexadded to the vial was adjusted to give a reading of1,000 kiloCounts on the Malvern. Before use, dis-tilled deionized water was filtered twice through afilter with an average pore size of 0.45 mm.

To validate the particle sizes as determined bythe Malvern, transmission electron micrographs(TEMs) were obtained for some of the latexes. Theparticular method used was a negative stainingtechnique using uranyl acetate as masking agent.Particles were deposited onto copper grids cov-ered with a thin carbon film before treatmentwith the masking agent. The particle sizes ob-tained with the Malvern compared well with theTEM results, although some shrinkage of the par-ticles occurred during the time the particles wereleft on the copper grid to dry.

RESULTS

The results will be discussed under two majorheadings, namely “Water-Soluble Initiator: Am-

monium Persulfate” and “Water-Insoluble Initia-tor: AIBN.” Each of these two sections will besubdivided into discussions of the responses forthe factorial experiments, namely particle size,polydispersity, and conversion.

Water-Soluble Initiator: Ammonium Persulfate

Table III contains the coefficients used to producethe response surface graphs in this section for thewater-soluble initiator, ammonium persulfate.

The effects of the individual factors and theircombinations are discussed in the next section.Initially, before monomer is added, most samplesof the epoxy-acrylate emulsion for both initiatortypes have a large polydispersity (between avalue of 0.6 and 1) and large average particle size(in the region of 1 mm and larger). As soon asmonomer is added, these large particles decreasein size and measured average particle sizes fall

Figure 2 Effect of salt in combination with a low levelof DMEA on particle size and polydispersity duringemulsion polymerization.

Figure 3 Response surface graphs of the effects of stirring speed, DMEA level, andthe addition of salt on polydispersity of the emulsion polymerization after 5 h ofpolymerization.

Figure 4 Effect of DMEA levels with salt on polydis-persity during emulsion copolymerization.


into the region of 120–300 nm, depending on theamount of dimethylaminoethanol DMEA used toneutralize the polyacrylate.

The only exception to this observation of initialparticle size before the addition of monomer iswhen a high DMEA level is combined with theaddition of salt. In such cases, measured particlesizes initially have a diameter of 630 nm.

Particle Size and Polydispersity

The response of the factorial experiment for thewater-soluble initiator, ammonium persulfate onparticle size, is given in Figure 1. Unless other-wise stated, all figures will be of experiments inwhich the stirring speed was held constant at 100revolutions per minute (rpm). A “system” in thecontext of this study will be defined as an emul-sion polymerization consisting of the ingredientsin Table II and conducted with a specific type ofinitiator.

From the response surface graphs shown inFigure 1, it is clear that the greatest influence onparticle size is represented by those graphs inwhich salt is an added factor. Particle size andpolydispersity both influence the latex rheologyand film formation of the endproduct. The great-est interaction is seen when salt is added in com-bination with variations in the level of DMEA.

The addition of salt causes agglomeration andeventual coagulation of the emulsion. This is mostpronounced when salt is added at a low level ofDMEA. Figure 2 illustrates the effect of this par-ticular combination on the particle size.

The combination of salt and a high DMEA levelaffects the particle size to a lesser extent than

does the combination of salt and a low DMEAlevel (refer to Fig. 2).

The effect of stirring speed, added salt, andvariation in the level of DMEA on the polydisper-sity of the emulsion can be seen in Figure 3. Theprominent effect of salt is evident from these re-sponse surface graphs.

Figure 4 shows the changes in polydispersitywhen salt is added to a system with a low DMEAlevel and a system with a high level of DMEA,respectively.

The effect on particle size and polydispersitywith constant stirring speed, no salt, and only thelevel of DMEA varied can be seen in Figure 5.Figure 6 shows the effect of this combination onpolydispersity.

In both cases, it can be seen that the change inparticle size (Fig. 5) and the polydispersity (Fig.6) remain virtually constant over time.

Of great interest is the virtual doubling in par-ticle size for the system in which a high level ofDMEA is used relative to a low level of DMEA(refer to Fig. 5).

Conversion

Figure 7 shows overall fractional conversion ofbutyl acrylate and styrene monomers to polymeras a response to the three factors investigated.

The significant effect on conversion exhibitedby the combination of DMEA and added salt, aswell as the effect of DMEA and salt individuallyon conversion can be seen in these response sur-face graphs (Fig. 7). Stirring speed alone has lessof an effect on conversion than in combination

Figure 5 Effect of DMEA level and no added salt onthe particle size of the emulsion polymerization.

Figure 6 Effect of DMEA levels and no added salt onthe polydispersity during emulsion polymerization.


with DMEA and added salt. Figure 8 shows theeffect that combinations of DMEA and salt haveon fractional conversion.

From Figure 8, it appears that addition of saltto the system is accompanied by low conversion.This is in contrast to the high conversion foundwhen no salt is added and the level of DMEA islow. The reason for this behavior should be soughtin the particular emulsion polymerization mech-anism operating under these conditions ratherthan concluding that the addition of salt is thereason for the observed reduction in conversion.

The combination of a high level of DMEA withthe addition of salt at any stirring speed levelshowed a less severe change in stability (changein particle size and polydispersity over time).

The stability of the emulsion containing addedsalt and a high level of DMEA did show a depen-dence on agitation to remain emulsified. As soonas agitation was stopped after polymerization, theemulsion flocculated. The floc was reemulsifiedwhen agitation was again applied later. Althoughthe exact mechanism still needs to be investi-gated, flocculation in this case is probably due toa bridging mechanism between the particles, giv-ing rise to a three-dimensional floc. When sam-ples were taken for particle size analysis, no floc-culation occurred because of a dilution effect bothin particle concentration and ionic strength of themedium.

The observed difference in rate of conversionbetween a system containing a high level of

Figure 7 Response surface graphs of the effects of DMEA level, the addition of salt,and stirring speed on overall fractional conversion of styrene and butyl acrylate to thepolymer after 6,000 s.

Figure 8 Effects of combinations of DMEA and salt levels on conversion. Stirringspeed had no real effect on the fractional conversion.


DMEA with added salt (Fig. 8) and one containingno salt was not as pronounced as in the case ofexperiments in which low levels of DMEA wereused.

The reason for these differences will be exam-ined in terms of differences in mechanisms oper-ating in the various emulsion polymerization sys-tems and will be reported on in a future publica-tion.21

Water-Insoluble Initiator: AIBN

The reason for choosing a water-insoluble initia-tor, such as AIBN, was to determine whetherthere are marked differences in the behavior ofemulsion copolymerization in terms of conver-sion, particle size, and polydispersity, comparedwith a system in which a water-soluble initiatorwas used. The coefficients describing the multiplelinear regression models for the various responsesare given in Table IV.

Particle Size and Polydispersity

In Figures 9 and 10, the effect of stirring speed inconjunction with added salt and variation inDMEA level shows a much greater effect on par-ticle size and polydispersity than was observedfor similar responses in the water-soluble initia-tor system.

Figure 11 shows the effect of combinations ofDMEA and salt levels on particle size for theAIBN initiated system. For the sake of compari-son, stirring speed is kept constant at 100 rpm.

Figure 12 shows the effect of the combinationsof DMEA and salt levels on the polydispersity forthe same experiments.

It can be seen from Figure 11 that the particlesize increases with time for the system in whichthe DMEA level is low and salt is added. Theemulsion eventually coagulates after a period of 3h. This behavior is similar to that observed for thewater-soluble initiator system. It is clear from

Table IV Estimates of Coefficients for Multiple Regression Model

Input Variables

Observed Response

Particle Size Polydispersity Conversion

A. Salt 414.54 0.07 20.065B. DMEA 42.85 0.029 20.0064C. Stirring speed 3.07 0.00039 0.0002AB 214.16 20.0025 0.000066AC 0.46 0.00017 20.000025BC 20.103 20.000014 21.67E-6Block 2167.8 20.064 20.032Constant 21077.72 20.057 0.848

Figure 9 Response surface graphs showing the effects of salt and DMEA levels andstirring speed on the particle size of an epoxy-acrylate emulsion after polymerizationwith a water-insoluble initiator system.


Figure 11 that a high level of DMEA increases theaverage particle size, compared with the systemin which a low level of DMEA was used. It is alsoevident from Figure 11 that a high level of DMEAin combination with added salt does not suffer thesame fate, namely coagulation, compared withthe system with a low level of DMEA.

Figure 12 compares the polydispersity for thevarious combinations of DMEA and salt levels ata constant stirring speed of 100 rpm. Of interestherein is the observation that a high DMEA levelwithout added salt yields the lowest polydisper-sity. It is also evident from Figure 12 that thepolydispersity of the system with a low level ofDMEA and added salt increases in time in thesame way as particle size. This would suggestthat the particle size and polydispersity of thesystem depend on the same factors for changes tobe observed.

An increase in particle size is seen when thestirring speed is increased from 100 rpm to 600rpm (see Fig. 13).

Figure 13 illustrates how the absence of addi-tional electrostatic stabilization combined withchanges in the velocity gradient and radius ofinteraction can influence the particle size.

Conversion

Figure 14 shows the effect of the various factorson conversion as a response for the AIBN initial-ized system.

Compared with the previous response surfacegraphs in Figures 9 and 10 for particle size andpolydispersity, the effect of stirring speed on con-version is much reduced.

Compared with the system in which a water-soluble initiator was used, we are now able to seemore clearly how the addition of salt affects theconversion of monomer to polymer with time (Fig.15). As in the case of reactions conducted with awater-soluble initiator, there is a major differencein the rate of conversion when the level of salt is

Figure 12 Effect of DMEA and salt levels on thepolydispersity during emulsion polymerization initi-ated by AIBN.

Figure 10 Response surface graphs of the effects of stirring speed, DMEA level andsalt on the polydispersity of an epoxy-acrylate emulsion after polymerization with awater-insoluble initiator.

Figure 11 Effect of DMEA and salt levels on theparticle size of emulsion polymerization initiated byAIBN.


varied from low to high for a system with a lowlevel of DMEA. The same change in the rate ofreaction is also observed for the system in whicha high level of DMEA was used with variation inthe level of salt. A reduction in the rate of reactionis observed (refer to Fig. 15). However, this reduc-tion in reaction rate is much less, compared to thecase in which a low level of DMEA is used.

DISCUSSION

To summarize the results of the various experi-ments conducted with the different initiators, theeffects of the various factors on the measuredresponses are given in terms of tables of F-ratios

for the different initiators. This statistic is ob-tained by dividing the sample variance by thepooled variance obtained for all responses. Aslong as the observations are normally distributed,the F-ratio provides a test to decide whether thevariance due to an observed response is signifi-cant or not. If the F-ratio is equal or smaller than1, the variance due to the observed response is notsignificant and the observation is probably due torandom variation. If the observed F-ratio is largerthan 1, the observed response is probably signif-icant.

It is well known that tertiary amines act asradical scavengers.22,23 The ability of theseamines to react with radical sources present dur-ing polymer degradation make them excellent

Figure 13 Effect of stirring speed on particle size for systems in which the DMEAlevel was varied.

Figure 14 Response surface graphs of the effects of stirring speed, DMEA level, andthe addition of salt on the conversion as a response for the water-insoluble initiatorsystem.


stabilizers. However, in the context of epoxy-ac-rylates, tertiary amines are used as catalysts forthe reaction of the epoxy groups and carboxylicacid groups during heat curing.13 They also de-crease the problem of odor associated with ammo-nia and other primary amines and yield a clearfilm upon heat curing. They may also aid in theeventual application viscosity of the emulsion andplay a part in the film formation process.

The use of a tertiary amine as a neutralizingagent for the carboxylic acid groups in the epoxy-acrylate system under investigation has causedsignificant changes in the conversion profiles ob-tained. This may be because of chain transferreactions occurring between monomeric free rad-icals and tertiary amines in the water phase andbetween growing polymer segments and tertiaryamines in the particles.

To clarify this observation, additional emulsionpolymerizations were conducted in which the ter-tiary amine was substituted with ammonium hy-droxide. The pH of the system neutralized by theammonium hydroxide was adjusted to reflect thatof a system with a low DMEA level (10 g) and thatof a high level of DMEA (40 g). The starting pH forthe lower level DMEA was 9.1. For the higherDMEA level, the starting pH was 11.

For the system in which the DMEA level wasadjusted, as in the previous experiments, withoutadded salt, particle size increased as before andthe same influence on conversion was observed asin previous experiments. The results of these ex-periments, in terms of average figures of particlesize and polydispersity are given in Table V. Con-version is taken after 6,000 s, one-third of thetotal time of polymerization.

It is evident from the results shown in Tables Vand VI that it is not the pH as such that affects

the reaction kinetics, but rather the type of amineused in neutralizing the polyelectrolyte. Of con-siderable interest is the effect that a high DMEAlevel (as in Table V) has on the particle size,compared with the use of ammonium hydroxideas a neutralizing amine (as in Table VI). It wouldseem that the type of counterion used in the neu-tralization of the polyacrylate influences the par-ticle size significantly.

From the results on particle sizes (Table V) andthe stability of latexes with high DMEA and saltlevels compared with those with low DMEA andsalt levels, it may be that the overriding mecha-nism of stabilization for such systems is stericstabilization.

The effect of salt on the outcome of the factorialexperiments seems to be especially important forthe experiments in which combinations of lowDMEA levels were used with added salt. The ef-fect was less clear with combinations of higherDMEA levels and added salt. The decrease inpolymerization rate observed with low DMEA lev-els and added salt seems to be linked to the fact

Figure 15 Effects of DMEA and salt levels on theconversion profile during emulsion polymerization ini-tiated by AIBN.

Table V Summary of Averaged ResultsObtained for Additional Epoxy-AcrylateEmulsion Polymerizations Using Low (10 g) andHigh (40 g) Levels of DMEA with No Added Saltand Stirring Speed Held Constant at 350 rpm

FactorsLow DMEA

ConcentrationHigh DMEA

Concentration

Particle size (nm) 144.67 248.055Polydispersity 0.453 0.278Conversion 0.76 0.36

Ammonium persulfate was used as initiator.

Table VI Summary of Results Obtained for anEpoxy-Acrylate Emulsion Polymerization inWhich pH Was Adjusted to Correspond to thepH of a Low Level of DMEA and a High Level ofDMEA, Respectively, While Containing NoAdded Salt and Keeping the Stirring SpeedConstant at 350 rpm

FactorsLow

pH SystemHigher

pH System

Particle size (nm) 143.0 171.52Polydispersity 0.319 0.279Conversion 0.96 0.94

Ammonium persulfate was used as initiator.


that significant coagulation occurred over timeduring polymerizations conducted under thesecombinations. This is seen in terms of observedchanges in the size of particles with time.

Reduction of the number of loci available forpolymerization should significantly affect the ki-netics of emulsion polymerization. Nomura24 dis-cussed the effect of particle size on the exit ofradicals from particles, as well as the contributionof termination to the overall kinetics of the poly-merization. Similar studies by Casey and col-leagues,25 Asua and colleagues26,27 and Nomura28

corroborate the idea that changes in the size andnumber of particles affect the mechanism of emul-sion polymerization in terms of exit and termina-tion of radicals, and are therefore the main rea-sons for the observed decrease in conversion rateobserved with variation in DMEA and salt levels.

From Tables VII and VIII, it is evident thatcertain factors are more pronounced when theyare compared for the two initiator types. Thegreater effect of stirring speed on the particle sizeand polydispersity for the AIBN-initiated systemprobably stems from the absence of additionalstabilization imparted by the ionic end groups ofthe ammonium persulfate-initiated system aspreviously discussed. In the case of the AIBN-

initiated system, this additional electrostatic sta-bility is not present, and therefore we see thegreater effect of stirring speed on particle size andpolydispersity (Table VIII).

The effect of DMEA as an individual factor onconversion can be seen in the magnitude of theF-ratios for conversion in Tables VII and VIII.DMEA may also affect the mechanism of emul-sion polymerization by acting as chain transferagent.

CONCLUSIONS

Based on the results of replicated factorial de-signed experiments conducted with a water-solu-ble initiator and a water-insoluble initiator, it canbe concluded that the type and concentration ofneutralizing amine influences the rate of mono-mer conversion probably through chain transfer,as well as the particle size. The combination ofsalt and a low level of tertiary amine as neutral-izing agent for the polyacrylate used to emulsifythe epoxy resin results in low conversion throughcoagulation of the emulsion that effectively re-duces the number of loci for polymerization.

Table VII F-Ratios for Individual Factors and Their Interactions Up to First Orderfor the System Initiated by the Water-Soluble Initiator, Ammonium Persulfate

Factor Particle Size Polydispersity Conversion

A. Stirring speed 0.00 0.87 4.2B. [DMEA] 2.9 7.3 34C. [Salt] 36 2.1 100AB 0.69 1.5 0.13AC 1 block 0.03 2.2 1.6BC 1 block 27 0.43 38Block 2.5 0.37 0.34

Table VIII F-Ratios for Individual Factors and Their Interactions Up to FirstOrder for the System Initiated by the Water-Insoluble Initiator, AIBN

Factor Particle Size Polydispersity Conversion

A. Stirring speed 1.6 6.7 1.2B. DMEA 3.8 7.8 28C. Salt 6.1 14 88AB 5.5 4.4 2.6AC 0.8 2.7 0.5BC 1.5 0.7 0.08Block 0.6 2.1 4.5


Combinations of high DMEA levels and addedsalt show less pronounced effects on particle size,polydispersity, and conversion, compared withsystems with a low level of DMEA. When compar-ing the effect of neutralizing amine type on par-ticle size and conversion, it is seen that the type ofamine plays a significant role. This observation isillustrated by the fact that a primary amine, suchas ammonium hydroxide, does not affect the rateof the reactions at pH values adjusted to corre-spond to those of the tertiary amine, dimethylath-anolamine, used in the factorial designs.

Although slight differences in the responsesare seen between the water-soluble initiator, am-monium persulfate, and the water-insoluble ini-tiator, AIBN, the basic trends in particle sizedevelopment, polydispersity, and conversion pro-files are similar for both systems.

REFERENCES

1. Allen, R. A.; Scott, L. W. U.S. Pat. 4,098,744 (1978).2. Allan, R. A.; Scott, L. W. U.S. Pat. 4,098,844 (1978).3. Evans, J. M.; Ting, V. M. U.S. Pat. 4,212,781

(1980).4. Georgalas, N.; Keaveney, W. P. U.S. Pat. 4,341,678

(1982).5. Dante, M. F.; Allan, R. A. U.S. Pat. 4,316,003

(1982).6. Richardson, F. Paint and Resin, October 1986, 15–25.7. Galgoci, E. C.; Kumar, P. C. Proceedings of the

Water-Borne Higher Solids Powder Coatings Sym-posium 1994, 21st, 54–64.

8. Bozzi, E. G. Modern Paints and Coatings 1977, 67,57–58.

9. Darwen, S. Eur Polym Paint and Colour J 1994,184, 65–67.

10. Ragupathy, K. Bull Electrochem 1986, 2, 361–362.11. Elmore, J. D.; Dickman, H. T. Proceedings of the

Water-Borne Higher Solids Coating Symposium1980, 7th 86–106.

12. Woo, J. T. K.; Ting, V.; Evans, J.; Marcinko, R.;Carlson, G.; Ortiz, C. J Coat Technol 1982, 54,41–55.

13. Isozaki, O. European Pat. 0,384,338 (1990).14. Kojima, S.; Moriga, T.; Gotoh, H. Proceedings of the

Water-Borne and Higher-Solids, and Powder Coat-ings Symposium, February 26–28, New Orleans,LA, USA, 146–152, 1993.

15. Hidalgo-Alvarez, R.; Martin, A.; Fernandez, A.;Bastos, D.; Martinez, F.; de las Nieves, F. J. AdvColloid Inter Sci 1996, 67, 1–118.

16. Atkins, P. W. Physical Chemistry, 3rd ed.; OxfordUniversity Press, Oxford, 1986.

17. El-Aasser, M. S. Advances in Emulsion Polymer-ization and Latex Technology, 27th Annual ShortCourse, Vol. 2; 1996, pp. 1–32.

18. Gilbert, R. G. Emulsion Polymerization, A Mecha-nistic Approach; 1995.

19. Morgan, E. Chemometrics: Experimental Design;John Wiley and Sons: Chichester, West Sussex,1995 (published on behalf of ACOL).

20. Khuri, A. I.; Cornell, J. A. 1996, Response Surfaces,Design and Analysis, 2nd ed., revised and expand-ed; Marcel Dekker, New York.

21. de Wet-Roos, D.; Knoetze, J. H.; Cooray, B. Sub-mitted.

22. Bamford, C. H.; Jenkins, A. D.; Johnston, R.;White, E. F. T. Transact Faraday Soc 1959, 55,168–178.

23. Bamford, C. H.; White, E. F. T. Transact FaradaySoc 1956, 52, 716–726.

24. Nomura, M. J Appl Polym Sci 1981, 26, 17–26.25. Casey, B. S.; Morrison, B. R.; Maxwell, I. A.; Gil-

bert, R. G.; Napper, D. H. J Polym Sci Part A:Polym Chem 1994, 32, 605–630.

26. Asua, J. M.; de La Cal, J. C. J Appl Polym Sci 1991,42, 1869–1877.

27. Asua, J. M.; Rodriguez, V. S.; Sudol, E. D.; El-Aasser, M. S. J Polym Sci: Part A: Polym Chem1989, 27, 3569–3587.

28. Nomura, M.; Ikoma, J.; Fujita, K. J Polym Sci: PartA: Polym Chem 1993, 31, 2103–2113.


Emulsion polymerization of an epoxy-acrylate emulsion stabilized with polyacrylate. i. influence of...

Documents

Transcript of Emulsion polymerization of an epoxy-acrylate emulsion stabilized with polyacrylate. i. influence of...